A year ago, I poked around Steam to see how many game developers were disclosing usage of Generative AI . It was around 1,000, which seemed like a lot to me at the time. If memory serves, that was about 1.1% of the entire Steam library, which has since seen 20,000+ more titles appear. I've been fol
Before, you knew that every single asset was placed by hand, and even if it was a prebuilt asset. A human was directly involved with every piece of artwork, dialogue, text, etc.
Now, you might come across dozens of random text documents or images that are seemingly and vaguely related to the story. How do I as the player know what’s actually relevant? Maybe the AI generated text sends me down a rabbit hole that has nothing to do with the game because it wasn’t proofread.
These were tasks that, even when menial, allowed for the artist to express themselves all the more. I’m imagining a painter being handed a premixed palette or a sculptor having someone apply the finishing touches for them.
It just feels like giving up at the finish line. Why do we need a bunch of unrelated text and images of the game stands fine without them?
You can make the same argument if the dev pays someone on fiver to type in some random crap and not proofread their work. But atleast it’s not AItm, right? Low effort is still low effort no matter the source.
Why do we need a bunch of unrelated text and images of the game stands fine without them?
Classic case of things you don’t notice because they’re there. But you will notice their absence.
Don’t put the blame on generative A.I for filler content. I’ve seen Loren ipsum as filler long before generative A.I was even invented. It’s always down to the devs to make sure what they put out is good.
But generative A.I absolutely make it easier to create the filler content, so you can focus on other aspects.
You’re putting out a lot of “what if”. If my grandma had wheels she’d be a bike and if my mom had balls she’d be my dad.
This has nothing to do with gen-A.I. so someone has lots of important lore that contradicts other lore? That’s their fault for misusing their tools. Not a fault of the tool.
It’s not like they would suddenly do a better job just because they didn’t use gen-A.I, they would still do a shitty job if they can’t pay attention to details.
It’s fine if you just don’t want any gen-A.I in what you purchase. That’s an opinion, a stance and I respect that. I just personally am not bothered if used correctly. I’m just complaining that their categories of gen-A.I doesn’t allow me to make that distinction.
At that point, the question a dev should be asking is, “Is it necessary to have filler text?” Taking your “background text that flashes onscreen” concept, if its just background noise anyway, why does it need to specifically be text? What do letters or characters add that couldn’t be achieved with glyphs or scribbles?
People often see AI as some shiny new tool to bring their visions to life, but the game design and storytelling techniques we’ve collectively learned over the past decades still apply. More ≠ better, and if it’s not meant to be consciously and overtly experienced by the player, what need is there to include it at all, AI-generated or otherwise?
We are not going to get anywhere debating a made up scenario. And what glyphs would or wouldn’t do instead of text is subjective. Maybe the game isn’t centered around ancient egypt?? I’m sure someone can find 1 million and 1 reasons as to why something might be text. But that’s not the topic.
I didn’t claim more = better. But I do think gen-AI has a place in some areas. You don’t have to agree.
In my view. It’s a tool. Tools can be misused. Tools can be useful.
I want to be clear that I used to have a similar mindset; I came around to the idea that if I’m a game dev (or any other type of creativity), and I’m looking for speedy shortcuts, I’m doing myself a disservice in two ways:
I’m not learning the necessary skills to improve my craft. It’s vibe coding in a different shirt, and skipping the struggle is skipping the parts that make me better.
I’m missing the forest for the trees. Extra details don’t make something better, and just because AI can do what I don’t want to do doesn’t mean that thing is worth doing. If I’m so disinterested in creating <insert thing>, why am I outsourcing it to a robot that doesn’t actually understand my art? Out of everyone, a dev should be the most interested in the details, and if it’s so unimportant that a robot can put in whatever, the end user likely wouldn’t have cared in the first place.
I agree that it’s a tool, and I agree that we aren’t likely to see eye to eye on this, but at the end of the day, I am wholly convinced that this tool is being used in the wrong place and in the wrong way. It doesn’t belong in creative endeavors.
I’m cool with you saying your final word on the matter, but I think we’ve beat this topic to death. You’ve been very courteous, and I appreciate that. Take care.
I’ll just reply with my 2 cents on your two points. This is the kind of topic where there isn’t a right or wrong. Just different opinions and people thinking differently. I hope you have a nice vacation and with that, here goes.
I know from personal experience that what you would like to do, does not often co-exist with the time you’re allowed to spent on it. I would have loved to refine so much, add better function, create a more robust framework to allow expansion of future functionality. But the machine is going out to a show in 4 weeks. It doesn’t need to be refined, it doesn’t need to be perfect. But it does need to work and be functional.
I did not use A.I for that project. Along with it, I had to create a UI that worked on a touch-screen. I had never done anything of the sorts before. I did study up and learn the basics in order to create the UI. Honestly, using A.I did not even cross my mind. It wasn’t a question of wanting or not wanting. It literally did not even occur to me that it’s something I could have used. But thinking back. If I had, maybe I would have saved a few hours and could have used that to finish some “almost done” functions.
Everyone is different, and I’m sure most devs are super interested in the details, they’re just, different details. The detail of certain text might not be as important to a dev as the detail of a certain framework and building it in a specific way to allow for future implementation of… whatever.
I’ll end it with, I’m sure there are devs of your description as well, that would rather die than let some machine taint their vision of what their project should be.
I suppose I see it from a very practical point of view. Rather than an artistic point of view. But I do understand the concern. And should quality start to suffer as a result, I’m sure my opinion would change.
So if that dev used AI to make various filler/background glyphs would it be okay? Because even scribbles take time to make, if an AI tool can do it quicker and its just background noise, is that okay?
Where is this imaginary line of acceptability? Its different for every person who enters these AI discussions. How about if the dev codes a tool that spits out procedural glyph assets, surely thats fine, but what is the real difference?
I think you’re getting caught up on the “thing.” My point wasn’t whether it should be glyphs or letters. Devs should be asking, at that point, why it’s necessary at all. Does it need to exist? If you’re considering AI just to generate “background noise,” is that noise really necessary in the first place? This step often happens naturally in human-derived work when we consider the work involved, but it must happen intentionally when you throw garbage-generators into the mix.
And no, I don’t think making glyphs via AI is okay, because now we’re in the realm of AI image generation, and that’s a giant unethical miasma. You ask what the difference is between a dev making their own script to general glyphs versus an AI, and that’s like asking what difference exists between a solar calculator and a data center run by Google. Both can tell you what 2+2 equals, but one is unnecessarily complex.
And then there’s the ethical considerations. Where did that AI model come from? How was it trained and developed? Whose work was used to derive that model? Who benefits from its public use?
AI simply does not belong in creative endeavors. People may have their own reasons for where they’ve drawn their lines, but that does not mean it is a mere matter of subjectivity, like choosing broccoli instead of carrots, or that they have a good basis for that decision.
I’m not even into this use of AI.
Before, you knew that every single asset was placed by hand, and even if it was a prebuilt asset. A human was directly involved with every piece of artwork, dialogue, text, etc.
Now, you might come across dozens of random text documents or images that are seemingly and vaguely related to the story. How do I as the player know what’s actually relevant? Maybe the AI generated text sends me down a rabbit hole that has nothing to do with the game because it wasn’t proofread.
These were tasks that, even when menial, allowed for the artist to express themselves all the more. I’m imagining a painter being handed a premixed palette or a sculptor having someone apply the finishing touches for them.
It just feels like giving up at the finish line. Why do we need a bunch of unrelated text and images of the game stands fine without them?
You can make the same argument if the dev pays someone on fiver to type in some random crap and not proofread their work. But atleast it’s not AItm, right? Low effort is still low effort no matter the source.
Classic case of things you don’t notice because they’re there. But you will notice their absence.
Don’t put the blame on generative A.I for filler content. I’ve seen Loren ipsum as filler long before generative A.I was even invented. It’s always down to the devs to make sure what they put out is good.
But generative A.I absolutely make it easier to create the filler content, so you can focus on other aspects.
You’re putting out a lot of “what if”. If my grandma had wheels she’d be a bike and if my mom had balls she’d be my dad.
This has nothing to do with gen-A.I. so someone has lots of important lore that contradicts other lore? That’s their fault for misusing their tools. Not a fault of the tool.
It’s not like they would suddenly do a better job just because they didn’t use gen-A.I, they would still do a shitty job if they can’t pay attention to details.
It’s fine if you just don’t want any gen-A.I in what you purchase. That’s an opinion, a stance and I respect that. I just personally am not bothered if used correctly. I’m just complaining that their categories of gen-A.I doesn’t allow me to make that distinction.
At that point, the question a dev should be asking is, “Is it necessary to have filler text?” Taking your “background text that flashes onscreen” concept, if its just background noise anyway, why does it need to specifically be text? What do letters or characters add that couldn’t be achieved with glyphs or scribbles?
People often see AI as some shiny new tool to bring their visions to life, but the game design and storytelling techniques we’ve collectively learned over the past decades still apply. More ≠ better, and if it’s not meant to be consciously and overtly experienced by the player, what need is there to include it at all, AI-generated or otherwise?
We are not going to get anywhere debating a made up scenario. And what glyphs would or wouldn’t do instead of text is subjective. Maybe the game isn’t centered around ancient egypt?? I’m sure someone can find 1 million and 1 reasons as to why something might be text. But that’s not the topic.
I didn’t claim more = better. But I do think gen-AI has a place in some areas. You don’t have to agree.
In my view. It’s a tool. Tools can be misused. Tools can be useful.
I want to be clear that I used to have a similar mindset; I came around to the idea that if I’m a game dev (or any other type of creativity), and I’m looking for speedy shortcuts, I’m doing myself a disservice in two ways:
I agree that it’s a tool, and I agree that we aren’t likely to see eye to eye on this, but at the end of the day, I am wholly convinced that this tool is being used in the wrong place and in the wrong way. It doesn’t belong in creative endeavors.
I’m cool with you saying your final word on the matter, but I think we’ve beat this topic to death. You’ve been very courteous, and I appreciate that. Take care.
I’ll just reply with my 2 cents on your two points. This is the kind of topic where there isn’t a right or wrong. Just different opinions and people thinking differently. I hope you have a nice vacation and with that, here goes.
I know from personal experience that what you would like to do, does not often co-exist with the time you’re allowed to spent on it. I would have loved to refine so much, add better function, create a more robust framework to allow expansion of future functionality. But the machine is going out to a show in 4 weeks. It doesn’t need to be refined, it doesn’t need to be perfect. But it does need to work and be functional.
I did not use A.I for that project. Along with it, I had to create a UI that worked on a touch-screen. I had never done anything of the sorts before. I did study up and learn the basics in order to create the UI. Honestly, using A.I did not even cross my mind. It wasn’t a question of wanting or not wanting. It literally did not even occur to me that it’s something I could have used. But thinking back. If I had, maybe I would have saved a few hours and could have used that to finish some “almost done” functions.
Everyone is different, and I’m sure most devs are super interested in the details, they’re just, different details. The detail of certain text might not be as important to a dev as the detail of a certain framework and building it in a specific way to allow for future implementation of… whatever.
I’ll end it with, I’m sure there are devs of your description as well, that would rather die than let some machine taint their vision of what their project should be.
I suppose I see it from a very practical point of view. Rather than an artistic point of view. But I do understand the concern. And should quality start to suffer as a result, I’m sure my opinion would change.
So if that dev used AI to make various filler/background glyphs would it be okay? Because even scribbles take time to make, if an AI tool can do it quicker and its just background noise, is that okay?
Where is this imaginary line of acceptability? Its different for every person who enters these AI discussions. How about if the dev codes a tool that spits out procedural glyph assets, surely thats fine, but what is the real difference?
I think you’re getting caught up on the “thing.” My point wasn’t whether it should be glyphs or letters. Devs should be asking, at that point, why it’s necessary at all. Does it need to exist? If you’re considering AI just to generate “background noise,” is that noise really necessary in the first place? This step often happens naturally in human-derived work when we consider the work involved, but it must happen intentionally when you throw garbage-generators into the mix.
And no, I don’t think making glyphs via AI is okay, because now we’re in the realm of AI image generation, and that’s a giant unethical miasma. You ask what the difference is between a dev making their own script to general glyphs versus an AI, and that’s like asking what difference exists between a solar calculator and a data center run by Google. Both can tell you what 2+2 equals, but one is unnecessarily complex.
And then there’s the ethical considerations. Where did that AI model come from? How was it trained and developed? Whose work was used to derive that model? Who benefits from its public use?
AI simply does not belong in creative endeavors. People may have their own reasons for where they’ve drawn their lines, but that does not mean it is a mere matter of subjectivity, like choosing broccoli instead of carrots, or that they have a good basis for that decision.